Table 3.
Comparison with related studies based at EDs.
| Study | Current Study | [17] | [18] | [19] | [20] |
|---|---|---|---|---|---|
| Study place | Taiwan | USA | Korea | USA | Germany |
| Study population | ED patients | ED patients | ED patients | ED patients | ED patients |
| Predicted outcome | High-risk adverse and critical care events (including hospitalization, sepsis or septic shock, ICU admission, in-hospital mortality, etc.) | Visualization of patients’ summarized data and flow | Visualization of patients’ summarized data and flow | Identification of altered mental status (AMS) | Suggested Diagnoses |
| AI/ML approach | 🗸 | N/A | N/A | 🗸 | 🗸 |
| Implementation | 🗸 | 🗸 | 🗸 | 🗸 | 🗸 |
| Real-time and individualized monitoring | 🗸 | 🗸 | 🗸 | N/A | N/A |
| Digital dashboard | 🗸 | 🗸 | 🗸 | N/A | N/A |
| AI/ML algorithm | 🗸 (including Random Forest, LightGBM, logistic regression, XGBoost, and MLP) |
N/A | N/A | 🗸 (including NLP, convolutional neural network) |
🗸 (No details reported) |
| Can adjust values to repeat predict | 🗸 | N/A | N/A | N/A | N/A |
| Notification alert | 🗸 | 🗸 | 🗸 | N/A | N/A |
| feature variable | 12–30 variables (including patients’ age, sex, Glasgow Coma Scale, vital signs, laboratory data, comorbidities, etc.) |
N/A | N/A | Text variable (clinical notes) |
Limited variables (including patient demographics, patient history, and information about current complaints) |
| Testing performance | AUC: 0.735–0.925 |
N/A | N/A | AUC: 0.985 |
Accuracy: (0.70–0.85) |
| Year | 2022 | 2017 | 2018 | 2019 | 2021 |
Note: AUC, the area under the ROC curve.